Implementation:Risingwavelabs Risingwave Handle Create Mv

Knowledge Sources	RisingWave RisingWave Docs
Domains	Streaming, Query_Processing, SQL_DDL
Last Updated	2026-02-09 07:00 GMT

Overview

Concrete tool for creating incrementally-maintained materialized views in RisingWave provided by the frontend SQL handler.

Description

The handle_create_mv function is the SQL DDL handler for CREATE MATERIALIZED VIEW statements. It binds the SQL query using the streaming binder, resolves dependent relations and UDFs, then delegates to handle_create_mv_bound which optimizes the query into a streaming plan and deploys it as a streaming job on compute nodes.

Usage

This handler is invoked when a user executes a CREATE MATERIALIZED VIEW ... AS SELECT statement. The resulting materialized view is continuously updated as new data arrives at its source relations.

Code Reference

Source Location

Repository: risingwave
File: src/frontend/src/handler/create_mv.rs
Lines: L158-186

Signature

pub async fn handle_create_mv(
    handler_args: HandlerArgs,
    if_not_exists: bool,
    name: ObjectName,
    query: Query,
    columns: Vec<Ident>,
    emit_mode: Option<EmitMode>,
) -> Result<RwPgResponse>

Import

// Internal to the frontend crate
use crate::handler::create_mv::handle_create_mv;

I/O Contract

Inputs

Name	Type	Required	Description
handler_args	HandlerArgs	Yes	Session context and handler metadata
if_not_exists	bool	Yes	Skip creation if MV already exists
name	ObjectName	Yes	Name of the materialized view
query	Query	Yes	SQL SELECT query defining the view
columns	Vec<Ident>	Yes	Optional column aliases (can be empty)
emit_mode	Option<EmitMode>	No	EMIT ON WINDOW CLOSE mode for windowed queries

Outputs

Name	Type	Description
Result<RwPgResponse>	RwPgResponse	Success response (StatementType::CREATE_MATERIALIZED_VIEW) or error
Streaming job	Side effect	Fragment graph deployed to compute nodes for incremental maintenance
Catalog entry	Side effect	Materialized view registered in system catalog

Usage Examples

Basic Aggregation MV

CREATE MATERIALIZED VIEW user_event_counts AS
SELECT
    user_id,
    event_type,
    COUNT(*) AS event_count,
    MAX(event_timestamp) AS last_event
FROM kafka_events
GROUP BY user_id, event_type;

Join with Temporal Filter

CREATE MATERIALIZED VIEW recent_orders AS
SELECT
    o.order_id,
    o.user_id,
    u.username,
    o.total_amount
FROM orders_source o
JOIN users_table u ON o.user_id = u.id
WHERE o.created_at > NOW() - INTERVAL '24 hours';

Related Pages

Implements Principle

Principle:Risingwavelabs_Risingwave_Materialized_View_Processing

Requires Environment

Environment:Risingwavelabs_Risingwave_Rust_Build_Environment

Uses Heuristic

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment